Quantitative multiwavelength constituent measurements using single-wavelength photon time-of-flight correction

A method to correct for spectral variations in multiwavelength responses using a photon time-of-flight measurement was investigated. A new approach is described where near-infrared spectra of various absorbing and scattering solutions are corrected for nonlinear variations in the measured attenuation. The spectral attenuation measured in a sample is related to the concentration of the analyte as well as the scattering from the medium. The dependence of the detected attenuation on the absorption and scattering at a single wavelength can be described by the steady-state diffusion approximation, With the use of the steady-state diffusion approximation, single-wavelength correction methodologies for multiwavelength responses were developed to refine constituent estimates in a scattering medium. The single-wavelength corrections examined the use of multiwavelength responses ratioed to the absorption and scattering determined at a single wavelength to reduce the scattering influence in the detected attenuations, The single-wavelength correction method improved constituent estimates in a variable medium. The correction utilized the absorption and scattering determined from processed time-resolved photon distributions at a single wavelength. Partial least-squares (PLS) regression was used to obtain constituent absorption estimates from diffuse transmittance spectra of turbid samples. A 70% improvement over PLS-processed spectra was obtained with a combined single-wavelength absorption and scattering correction. In comparison to steady-state multiwavelength correction methods such as wavelength normalization, derivative spectroscopy, and multiplicative signal correction, the results demonstrate a significant improvement in estimates of the dye concentration. Results suggest a practical method to correct for scatter variations in multiwavelength responses for routine spectroscopic chemical analysis of turbid samples.